And do you think Taleb is speaking in a specific, precise, highly-technical, highly unusual definition of ‘complex’ rarely found outside of computer science circles and extremely niche interest groups like LessWrong? Or do you think he is speaking in the usual colloquial sense of ‘complex’ which everyone understands and under which physics is indeed extremely complex and difficult?
Chaos theory is a part of physics that deals with complexity that Taleb would probably call complex. On the other hand I don’t think that Taleb would call classical Newtonian physics complex.
Or consider the modern economy—when I contemplate a wireless mouse, it seems vastly more complex than the problem it solves (a mouse which requires a wire), yet, the wireless mouse still works & is nice to have.
I don’t think that Taleb would be a fan of wireless mouses. A wireless mouse has the failure mode of the battery dying. A wired mouse doesn’t have that problem because it’s less complex. Look at his website. Despite various people offering to him to do the necessary work, he doesn’t use a content management system like Wordpress.
When it comes to software design Taleb probably favors the 37 signals philosophy. It tries to solve problems in a way that’s as simple as possible instead of being complex.
Take tax law as another example. Politicians want to encourage certain behavior so they write an exception into the tax law that people who engage into that behavior have to pay less taxes. It adds complexity to the system even if you get some people to shift their behavior in the right direction. The result is that developed countries have very complex tax laws that nobody really understands.
Taleb would recommend to make the tax law simpler by not trying to form the law in a way that fixes every single exception and encourages people to engage in specific actions. Whenever congress passes a tax law the tax code shouldn’t grow in size but shrink.
Changes in tax law should not increase it’s Kolmogorov complexity but reduce it. Washington politicians don’t understand it. Even economics professors like our Hanson don’t.
Low Kolmogorov complexity tax law would also be easier to understand in the more colloquial sense of ‘complex’.
Chaos theory is a part of physics that deals with complexity that Taleb would probably call complex. On the other hand I don’t think that Taleb would call classical Newtonian physics complex.
These two statements are contradictory. Lots of Newtonian physics problems are chaotic- in fact Poincarre developed the early ideas that became chaos theory to deal with the three body problem.
Chaos theory is a part of physics that deals with complexity that Taleb would probably call complex.
Would he?
I don’t think that Taleb would be a fan of wireless mouses.
More the worse for him. The markets have spoken and they like their wireless mice. Wirelessness, for all the added complexity—complexity which is much more complex than the problem it solves—seems to be carrying its weight, contra the Taleb quote.
Changes in tax law should not increase it’s Kolmogorov complexity but reduce it. Low Kolmogorov complexity tax law would also be easier to understand in the more colloquial sense of ‘complex’.
I doubt that. Here is where using a complex and not widely-understood theory can bite you: the shorter the code, the more computation you expect it to use to yield the specified results. The shortest possible code for anything reasonably complex will take huge amounts of computing power. (How much computation to simulate our universe up to the present moment using the minimal encoding of the Standard Model + initial conditions?) Any shorter version of the current tax code which has the same meaning will be much harder to understand than a version which redundantly spells out details and common cases for you. And if you change the actual meaning of the tax code, well, then you run into the public choice issues which made the meaning what it is now...
(It would be better to talk about logical depth and sophistication than simple uncomputable Kolmogorov complexity, but I believe even less that that is what Taleb meant.)
Washington politicians don’t understand it. Even economics professors like our Hanson don’t.
Why should they? Do we expect physicists to understand each and every area of physics? Specialization is one of the defining principles of modern societies.
And do you think Taleb is speaking in a specific, precise, highly-technical, highly unusual definition of ‘complex’ rarely found outside of computer science circles and extremely niche interest groups like LessWrong? Or do you think he is speaking in the usual colloquial sense of ‘complex’ which everyone understands and under which physics is indeed extremely complex and difficult?
Chaos theory is a part of physics that deals with complexity that Taleb would probably call complex. On the other hand I don’t think that Taleb would call classical Newtonian physics complex.
I don’t think that Taleb would be a fan of wireless mouses. A wireless mouse has the failure mode of the battery dying. A wired mouse doesn’t have that problem because it’s less complex. Look at his website. Despite various people offering to him to do the necessary work, he doesn’t use a content management system like Wordpress.
When it comes to software design Taleb probably favors the 37 signals philosophy. It tries to solve problems in a way that’s as simple as possible instead of being complex.
Take tax law as another example. Politicians want to encourage certain behavior so they write an exception into the tax law that people who engage into that behavior have to pay less taxes. It adds complexity to the system even if you get some people to shift their behavior in the right direction. The result is that developed countries have very complex tax laws that nobody really understands.
Taleb would recommend to make the tax law simpler by not trying to form the law in a way that fixes every single exception and encourages people to engage in specific actions. Whenever congress passes a tax law the tax code shouldn’t grow in size but shrink.
Changes in tax law should not increase it’s Kolmogorov complexity but reduce it. Washington politicians don’t understand it. Even economics professors like our Hanson don’t.
Low Kolmogorov complexity tax law would also be easier to understand in the more colloquial sense of ‘complex’.
These two statements are contradictory. Lots of Newtonian physics problems are chaotic- in fact Poincarre developed the early ideas that became chaos theory to deal with the three body problem.
Would he?
More the worse for him. The markets have spoken and they like their wireless mice. Wirelessness, for all the added complexity—complexity which is much more complex than the problem it solves—seems to be carrying its weight, contra the Taleb quote.
I doubt that. Here is where using a complex and not widely-understood theory can bite you: the shorter the code, the more computation you expect it to use to yield the specified results. The shortest possible code for anything reasonably complex will take huge amounts of computing power. (How much computation to simulate our universe up to the present moment using the minimal encoding of the Standard Model + initial conditions?) Any shorter version of the current tax code which has the same meaning will be much harder to understand than a version which redundantly spells out details and common cases for you. And if you change the actual meaning of the tax code, well, then you run into the public choice issues which made the meaning what it is now...
(It would be better to talk about logical depth and sophistication than simple uncomputable Kolmogorov complexity, but I believe even less that that is what Taleb meant.)
Why should they? Do we expect physicists to understand each and every area of physics? Specialization is one of the defining principles of modern societies.